The primary aims of the proposed project are to (1) Characterize structural and functional connectivity of the speech and language networks in minimally verbal children with autism spectrum disorder (ASD); and (2) Investigate changes in this connectivity resulting from intervention, specifically the Auditory Motor Mapping Training (AMMT) of Schlaug and colleagues (ACE Project 1). Two neuroimaging studies involving children with ASD and neurologically normal control subjects will be performed. In Study 1, we will perform high-resolution magnetic resonance imaging (MRI) scans of the brains of each participant from four groups of adolescents (ages 14-17): (1) 25 minimally verbal ASD adolescent, (2) 25 verbal but language impaired ASD adolescents, (3) 25 normal language ASD adolescents, and (4) 25 age- and gender-matched neurologically normal control adolescents. We will quantify structural anatomical connectivity based on Diffusion Tensor Imaging (DTI) data and functional connectivity based on resting state functional connectivity MRI, and we will use these measures to test hypotheses concerning connectivity deficits in the speech networks of adolescents with ASD. Particular attention will be paid to the pathway between left supplementary motor area (SMA) and left ventral premotor cortex (vPMC), which is hypothesized to play a key role in initiation of verbal output and which has been shown to be impaired in ASD. Furthermore, language-related hypotheses,including predictions regarding Inter-hemispheric connectivity of language areas BA 45 and BA 9 (in keeping with the aims of Project IV of this ACE application) will be investigated. Exploratory analyses of brain connectivity in ASD will also be performed on these data. In Study 2, we will investigate the effects of AMMT therapy on brain connectivity in ASD by performing the same scans used in Study 1 on 30 minimally verbal children with ASD (ages 6-10) before and after AMMT training performed as part of Project 1 of this ACE application. A central hypothesis tested in Study 2 is that AMMT will increase the functional connectivity and structual integrity of the white matter between left SMA and vPMC and that these variables, measured at the baseline scan, will also predict the response to AMMT training.